Hi!  Need some help.  I have am doing some data cleaning on a list of CHARACTER variables (which I will later turn to numeric- but do not want to yet).   The expected range of raw data: missing, 0 to 10, and some random character alpha values.  I need the expected range of cleaned numeric variable: missing, 0 to 10     I am trying to do this with an array, but not working correctly.  Basically, if the variable is between 1-10 then leave it as is.  If it is not, then set to missing.   data rr.recovperson2006 (keep = VHFUHFCommACFT VHFUHFCommShip VHFUHFNavACFT VHFUHFNavShip gpsACFT gpsSHIP) ;     set rr2006labels;         array change _character_;         do over change;               if (change>=0 and change<=10) then change=change;                 else change=' ';         end; run;   Help is appreciated!!   ... View more

Hi!  I have a variable (concurrentcrg)..and I need to flag all the observations with values that start with 5 but do not end in 1.     Here the line of code that did not work...   if concurrentcrg in: ('5') and concurrentcrg ne '%1' then status='Uncontrolled';   Help is appreciated!     ... View more

Hi all!  Is there a more efficient way to write this sql join?  The tables I am querrying are not that large, but the processing time is taking forever.  Basically, I have unique ids in table A (patid_copy).  I want to pull all records from b,c,d,e,f - that have matching id (patid_copy=patuniq).  Thanks! proc sql; create table  asthma.missing_sidr as select a.*,b.*, c.*, d.*,e.*,f.* from patid as a ,sidr.mdrsidrfy10 as b ,sidr.mdrsidrfy11 as c ,sidr.mdrsidrfy12 as d ,sidr.mdrsidrfy13 as e ,sidr.mdrsidrfy14 as f where a.patid_copy= b.patuniq or a.patid_copy= c.patuniq or a.patid_copy= d.patuniq or a.patid_copy= e.patuniq or a.patid_copy= f.patuniq; quit; ... View more

Hi!  I'm looking for help with this array.  I want to check multiple variables (Secondarydiagnosis1-20, Procedure1-20, and ProcedureHCPCS_1-13)   for a number starting with '493'.  If it begins with '493' then new variable asthma=1.  I submitted this code but it is taking a long time to process, so I'm sure there is an error!  Help is appreciated!     data asthma2.asthma_only; set asthma2.fyall_EXCL_F; asthma=0; array  dx(*)secondarydiagnosis2-secondarydiagnosis20; do i= 1 to dim(dx); if dx(i) =: '493' then asthma=1; array proc(*) procedure1-procedure20; do i=1 to dim(proc); if proc(i)=:'493' then asthma=1; array HCPCS(*) ProcedureHCPCS_1-ProcedureHCPCS_13; do i=1 to dim(HCPCS); if hcpcs(i)=:'493' then asthma=1; end; end; end; run; ... View more

Hi!  How do i change a variable (X) with a date format mmddyy10. to a SAS date?     ... View more

I have a dataset composed of 2.5 million subjects that contain case (study=1) and controls (study=0) . I need SAS coding to create a new dataset composed of all the cases plus five randomly selected controls for each case, matched by date, sex, and branch of service.   Control should only be used once (if already assigned to a case, cannot be assigned to another case).  This is my attempt, but not correct...   data tmp; set exp2.case exp2.control; rand_num=rand("uniform"); run; *merge case/ control tables and assign random number;     proc sort data=tmp; by ssn study rand_num; run; * 'study' variable is 1=case 0=control;   proc rank data=tmp out=tempR; by ssn study; var rand_num; run; THIS CHANGED ALL RAND_NUM to ‘1’ WHICH I DON”T THINK IS CORRECT?   proc sql; create table exp2.match as select a.ssn as e_ssn, a.service_branch as e_service_branch, a.sex as e_sex, a.service_entry_date as e_service_entry_date, a.study, b.ssn as ne_ssn, b.service_branch as ne_service_branch, b.sex as ne_sex, b.service_entry_date as ne_service_entry_date, b.study from tempr as a inner join tempr as b      on( a.service_entry_date between b.service_entry_date and b.service_entry_date_low) and a.service_branch=b.service_branch and a.sex=b.sex and a.rand_num=ceil(b.rand_num/5) where a.study and not b.study; quit;   Data looks like this + 2 million encounters Obs SSN study service_branch sex service_entry_date service_entry_date_low rand_num 1 1 0 A MALE 05/10/2010 05/10/2009 1 2 2 1 A MALE 11/20/2006   1 3 3 0 M MALE 10/07/2011 10/07/2010 1 4 4 1 A MALE 06/22/2010   1 5 5 0 M MALE 10/23/2007 10/23/2006 1 6 6 0 M MALE 07/08/2013 07/08/2012 1 7 7 0 F FEMALE 01/16/2007 01/16/2006 1 8 8 1 A MALE 08/03/2006   1 9 9 0 F MALE 01/20/2009 01/21/2008 1 10 10 0 A MALE 01/06/2009 01/07/2008 1 11 11 0 A MALE 03/18/2009 03/18/2008 1 12 12 1 A FEMALE 04/19/2007 . 1 13 13 0 F MALE 02/18/2014 02/18/2013 1 14 14 0 F FEMALE 09/04/2007 09/04/2006 1 15 15 0 A MALE 08/18/2014 08/18/2013 1 16 16 0 N FEMALE 11/18/2013 11/18/2012 1 17 17 1 N FEMALE 07/08/2014   1 18 18 0 N MALE 03/11/2014 03/11/2013 1 19 19 0 N MALE 09/17/2013 09/17/2012 1 20 20 0 N MALE 03/13/2014 03/13/2013 1     ... View more

Hi!  I have this logistic regression code w/ Odd Ratio Plot.   I need to  output/ print just the Odds Ratio Plot in a jpg, tiff, or psd w/  300 dpi for use in a manuscript. I am really unclear how to use ODS...  I use ods rtf  a lot for just writing in word...but the resolution is low.  What are my other options?   proc logistic data=asthma.edip_vitals simple plots(only)=oddsratio (cldisplay=line                         order=ascending                         type=horizontalstat                         range=(0.5,1.5));title 'Family Structure Effects on ED/Inpatient Utilization'; class household(ref='MM') structure (ref='Married Male')marital_status1 (ref='Married') sponsex1 (ref='Male') mult_dependents (ref='No') new_race1 (ref='White')pay_grade1 (ref='Enlisted: E1-E5')branch (ref='Army') agegroup (ref='0') education (ref='HS or less')sex_d (ref='0')/ param=ref; model utilization (event='1')=structure agegroup mult_dependents new_race1 pay_grade1 education; oddsratio structure; run;     Help is appreciated!!   ... View more